Forty Pages of Charts and Eleven Rows of Data

LoL Park sits inside the Gran Seoul building, and on the night of the upper-bracket series, the third-floor corridor was as hot as a server room. I was in the press area, holding a forty-page deck that a member of a coaching staff had handed me before the match. A traditional sports writer would say "before kickoff"; there is no ball here, only five chairs and one large screen. Forty pages. Radar charts for every lane. Heat maps of vision-control zones. A champion-pool depth matrix, split into four groups and shaded in four tones. A regression line predicting win rate against the timing of the first objective. The presentation was so beautiful that I nearly forgot my cheapest and most useful professional reflex: always flip to the appendix before believing anything. The appendix contained eleven rows of data. Eleven. Three of them misspelled player names. I sat still for about two minutes in the noise of the corridor, and then I understood something I had suspected for two years without having evidence pretty enough to say out loud: that document did not exist to analyze. It existed to look like analysis. And that night, the team that handed it to me lost the series. Not because of eleven rows of data. Because they believed in forty pages of charts. To understand why that story is not a trivial anecdote, we have to go back to January 2026, when the LCK and the LPL both switched to Fearless Draft in BO3 and BO5 alike. The format can be described in one sentence: a champion picked once is locked for the rest of the series. No refunds. No second thoughts. No negotiation. Simple arithmetic: a BO5 that runs the full five games consumes up to fifty ban-pick slots, twenty-five bans and twenty-five picks. League of Legends has more than one hundred and seventy champions. That still sounds roomy, until you remember that within a given meta only roughly sixty to seventy champions are genuinely playable at the top level. A single Fearless BO5 can burn through nearly half of that playable pool. Not in a Worlds final. In the group stage. The analytical class responded almost in unison, and that is when my ear started to itch. Everyone said the same sentence: Fearless Draft rewards teams with deep champion pools. Broadcasters built graphics about pool depth. Podcasts opened with the phrase. Stats sites began selling data packages on champion-diversity indices. Within three weeks, "champion pool depth" had become an asset class that could be priced, traded, and shown off in press conferences. I have sat in LCK arenas and LCK press rooms long enough to extract one rule: when an entire industry agrees too quickly on a concept, that concept is usually undefined. Champion pool depth is such a concept. Everyone uses it. Nobody can measure it. I once sat with a data analyst from an LCK team in a cafe in Gangnam, and he walked me through how he built his pool-depth index. Step one: pull twelve months of pick history. Step two: count distinct champions. Step three: weight by recency. Step four: normalize to a hundred-point scale. Step five: draw the radar. I asked him one question: if Fearless Draft means each champion is played only once in a series, how much of your sample is left? He was quiet for about ten seconds, then said: "Not enough." That is the entire problem, packed into two words. Champion pool depth is not a measurement. It is a story told with numbers. Suppose a mid-lane champion is picked seven times in a split and wins four. Win rate: 57.1 percent. The report prints 57.1 to one decimal place, with a green cell beside it. The ninety-five percent confidence interval for a proportion with n equal to seven and p around 0.57 stretches from roughly 0.25 to roughly 0.86. In the language of people who actually work with data: the same dataset is compatible with the hypothesis that this champion is terrible, and equally compatible with the hypothesis that it is broken. The report does not say that. The report prints 57.1 and colors it green. Multiply that by the three hundred pages of statistics this industry publishes every split, and you have a mountain of conclusions built on confidence intervals as wide as a map. The statistical problem, though, is only the shell. The meat is somewhere else. Champion pool depth comes in two kinds, and the industry only measures the useless one. I call them nominal depth and functional depth. Nominal depth is the number of distinct champions a player has competed with across a career. That is the number you see on stats sites. It is easy to count, easy to chart, easy to sell. Functional depth is the number of distinct tactical roles that player can execute at international level, while being targeted in the draft, with no preparation time. Nobody measures that number, because it requires reviewing footage at a frequency no data department has the patience for. A mid laner has twelve champions in his competitive record. Of those twelve, nine are vision-control mages, two are melee assassins, one is a long-range mage. Nominal depth: twelve. Functional depth: about two, maybe three, depending on the patch. The opponent only has to ban out two core mages and take a third in the first rotation, and the whole twelve-champion pool collapses in forty-five seconds. Here is the detail the radar chart never draws: most of the champions in a professional's nominal pool are champions that have never been banned. They play them in scrims, in practice, in matches nobody watches. Nominal depth is built out of occasions no one bothered to challenge. When someone finally challenges it, it usually lasts one game. Chovy is famous for a wide mid-lane pool, but the interesting thing about him is not the count. It is that the champions in that pool do not serve the same purpose. Faker, at twenty-nine, is the mirror case in an interesting way: he does not have the most champions, but he is one of the few who can change tactical role without changing the team's tempo. Both cases say the same thing: what gets measured and what matters are two different objects. The transfer market this season reflects that illusion directly. Across the last two transfer windows, a player's price has been set largely by the number of champions on his record. Teams pay for nominal flexibility: someone who plays three roles, someone who can play both assassins and mages, someone with fifteen champions in his competitive history. I have read contracts in which a player's bonus is tied to a flexibility index built by the team itself, using exactly the champion-counting method I just dissected. It is a perfect loop. The team builds the index. The team uses the index to price the player. The player knows the index exists, so he grinds new champions in solo queue. The index rises. The team pays more. And nobody in that loop checks whether the newly ground champions actually work when the draft targets him. I am not saying flexibility has no value. I am saying the market is paying for the kind of flexibility that is easy to measure instead of the kind that matters, and in a transfer window that is precisely the most expensive category of mistake, because it rides along on a two- or three-year contract. The real variable of Fearless Draft is not the champion pool. It is decision speed when the map is running dry. Under the old format, the draft was a two-way information game with forty seconds of thinking time per turn. Under Fearless, it becomes a resource-depletion game. Every pick both removes a champion from the opponent and removes one from you, from the next game, from game four, from game five. You are banning yourself twenty minutes into the future. Based on my experience watching matches at LoL Park, I sit close enough to the coaching desks to see what the broadcast never shows. In game four of a BO5, the coach is no longer reading an analysis deck. He is reading players' faces. The forty-page document lies there on the table, pages face down. Nobody flips it back open when the draft clock has fifteen seconds left. A team's real capacity in Fearless Draft is the ability to re-plan in forty-five seconds, and that lives in no dataset. It lives elsewhere. It lives in the moment a player who has been silent all series suddenly says one short sentence and the whole team follows. It lives in a coach daring to discard the plan he drew ten seconds earlier. It lives in the courage to accept a bad but familiar position instead of a beautiful one nobody has ever played. And this is where I have to say plainly something my own profession does not like hearing. Seoul that year did not rebel; it merely showed that tactics are written after the match is over. I have sat in hundreds of press conferences. I can predict the structure of ninety percent of them. The coach will say his team deliberately baited the opponent into a specific pick, that it was part of the plan, that they had prepared for this situation since the start of the week. Then I go home, open the replay, and I see that in the third minute his team placed a ward in a spot completely inconsistent with the plan described. I see his team take forty seconds to react to a lane swap the plan should have accounted for. The plan did not exist before the match. The plan was written in the press room, from a result that already existed, by a man who needed to protect his chair. That is not entirely a lie. It is worse. It is storytelling. And it is this storytelling trade that produced forty pages of charts from eleven rows of data, because once you accept that the story matters more than the measurement, replacing the measurement with a picture is only one step away. To be fair, I have to put on the table the number that argues against me. In the personal tracking sheet I kept through the 2026 LCK season, the team with the highest champion-concentration index, meaning the narrowest pool by conventional measurement, was among the teams that went furthest. If pool depth were the prerequisite the analytical consensus claims, that number should not exist. A team without a deep pool should have been eliminated in game four of its first series. That number does not overturn my argument, but it forces me to be more precise about it. When I reopened that team's footage, their narrow pool had a property the champion count cannot capture: the champions in it did not overlap in function. Each served a different purpose, and more importantly, each belonged to a functional group the opponent could not ban out simultaneously. They did not have ten champions for one job. They had three champions for three jobs, and all three were good enough. In other words, they were poor nominally and rich functionally. Exactly the paradox the radar chart cannot draw. And I admit it: if I had counted champions like everyone else, I would have misread that team all season. The esports analysis industry has built itself a production line. Take raw data. Run it through a few charting tools. Add a layer of professional language. Ship a document that looks authoritative. The line runs so fast that nobody manages to install a check at the end. And because there is no check, it never reports an error. Empty input, and the output is still a forty-page document with a hard cover and a signature. Teams do not die from a lack of talent. They die from believing in the meta, in scrims, in spreadsheets, more than in the hands trembling on their own keyboards. I saw that in another league, another sport, years ago, and I still remember the feeling of realizing it. Football walked this road about fifteen years ahead of us. There, distance covered and sprint counts were packaged as effort metrics. A midfielder who runs thirteen kilometers in a match gets highlighted on a television graphic. Nobody asks what those thirteen kilometers were for. You can run thirteen kilometers without touching the ball at the right moment at all, and still receive a green cell. The whole world chants pressing, and all I see is a crowd chasing the ball as if it were the truth. Esports is walking that same road, only faster. We have more data than football, we collect it more easily, and we chart it faster. Which means we can manufacture wrong conclusions at a speed football needed fifteen years to reach. Where can I be wrong? There are four possibilities, and I list them seriously because I do not want to become the mirror image of the people I am criticizing. Possibility one: Fearless Draft is temporary. Riot Games may revert after a season, or confine it to one segment. If so, the entire pool-depth problem disappears with the format, and I will be the man who wrote a long piece about a stillborn concept. Possibility two: one season is not a sample. I am leaning on a single year of observation, and I know perfectly well that this violates the very standard I just applied to everyone else. If the 2026 season proves that teams with deep nominal pools dominate, my functional-depth argument will have to shrink into a local phenomenon rather than a rule. Possibility three: I am part of the problem. I have published analyses built on far too little data. I have written lines about a team's tactical identity after watching exactly three of their games. Dig through my archive from three years ago and apply the standard I am applying to this industry, and you will find plenty of material to bury me. Possibility four, and this is the frightening one: perhaps pool depth is a real variable, merely measured with a bad instrument. If so, my argument is not a revelation. It is just the obvious statement that measurement is hard. And an obvious statement does not deserve two thousand words of anyone's time. I choose to live with those four possibilities rather than pretend they do not exist. If forced to make a verifiable judgment, I would put it this way. By the end of the 2026 season, I expect fewer than five teams worldwide to be capable of running a full Fearless BO5 without repeating a single tactical function twice at a level good enough to win. If that number holds, what decides titles over the next two years will not be the champion pool, but the capacity to decide under the pressure of depletion, something with no index, no chart, and therefore no one selling it to you. And if I am wrong, I will be the first to sit down and rewrite it, with this same pen. And if a forty-page document can be written out of eleven rows of data, then of the analyses you clicked on today, how many were the same?